hollalinux
/
linux-sg2042
mirror of https://github.com/sophgo/linux-riscv.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

debugfs: add skip_reset_on_timeout option 

To be able to debug long-running CS better, without changing the
userspace code, we are adding a new option through debugfs interface
to skip the reset of the device in case of CS timeout.

Signed-off-by: Yuri Nudelman <ynudelman@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
This commit is contained in:
Yuri Nudelman 2021-06-13 09:22:20 +03:00 committed by Oded Gabbay
parent 38e19d0b87
commit 4d041216c8
4 changed files with 17 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
Documentation/ABI/testing/debugfs-driver-habanalabs
View File

@ -207,6 +207,14 @@ Contact: ogabbay@kernel.org
Description: Sets the PCI power state. Valid values are "1" for D0 and "2"
for D3Hot
What: /sys/kernel/debug/habanalabs/hl<n>/skip_reset_on_timeout
Date: Jun 2021
KernelVersion: 5.13
Contact: ynudelman@habana.ai
Description: Sets the skip reset on timeout option for the device. Value of
"0" means device will be reset in case some CS has timed out,
otherwise it will not be reset.
What: /sys/kernel/debug/habanalabs/hl<n>/stop_on_err
Date: Mar 2020
KernelVersion: 5.6

1
drivers/misc/habanalabs/common/command_submission.c
View File

@ -663,6 +663,7 @@ static int allocate_cs(struct hl_device *hdev, struct hl_ctx *ctx,
cs->timestamp = !!(flags & HL_CS_FLAGS_TIMESTAMP);
cs->timeout_jiffies = timeout;
cs->skip_reset_on_timeout =
hdev->skip_reset_on_timeout ||
!!(flags & HL_CS_FLAGS_SKIP_RESET_ON_TIMEOUT);
cs->submission_time_jiffies = jiffies;
INIT_LIST_HEAD(&cs->job_list);

5
drivers/misc/habanalabs/common/debugfs.c
View File

@ -1278,6 +1278,11 @@ void hl_debugfs_add_device(struct hl_device *hdev)
dev_entry->root,
&dev_entry->blob_desc);
debugfs_create_x8("skip_reset_on_timeout",
0644,
dev_entry->root,
&hdev->skip_reset_on_timeout);
for (i = 0, entry = dev_entry->entry_arr ; i < count ; i++, entry++) {
debugfs_create_file(hl_debugfs_list[i].name,
0444,

3
drivers/misc/habanalabs/common/habanalabs.h
View File

@ -2191,6 +2191,8 @@ struct hl_mmu_funcs {
* @supports_staged_submission: true if staged submissions are supported
* @curr_reset_cause: saves an enumerated reset cause when a hard reset is
* triggered, and cleared after it is shared with preboot.
* @skip_reset_on_timeout: Skip device reset if CS has timed out, wait for it to
* complete instead.
*/
struct hl_device {
struct pci_dev *pdev;
@ -2305,6 +2307,7 @@ struct hl_device {
u8 device_fini_pending;
u8 supports_staged_submission;
u8 curr_reset_cause;
u8 skip_reset_on_timeout;
/* Parameters for bring-up */
u64 nic_ports_mask;